Tell me why: Visual foundation models as self-explainable classifiers
This work addresses interpretability for critical applications in AI, offering an efficient solution with a lightweight head, but it is incremental as it builds on existing VFM and SEM approaches.
The paper tackled the problem of interpretability in visual foundation models by proposing ProtoFM, a self-explainable classifier that combines VFMs with a prototypical architecture and specialized training objectives, achieving competitive classification performance and outperforming existing models on interpretability metrics.
Visual foundation models (VFMs) have become increasingly popular due to their state-of-the-art performance. However, interpretability remains crucial for critical applications. In this sense, self-explainable models (SEM) aim to provide interpretable classifiers that decompose predictions into a weighted sum of interpretable concepts. Despite their promise, recent studies have shown that these explanations often lack faithfulness. In this work, we combine VFMs with a novel prototypical architecture and specialized training objectives. By training only a lightweight head (approximately 1M parameters) on top of frozen VFMs, our approach (ProtoFM) offers an efficient and interpretable solution. Evaluations demonstrate that our approach achieves competitive classification performance while outperforming existing models across a range of interpretability metrics derived from the literature. Code is available at https://github.com/hturbe/proto-fm.